Mathematical Framework for Multi-Camera Cooperative Scene Interpretation
Sebastian Grünwedel · Ghent University Academic Bibliography (Ghent University) · 2010
Video surveillance is important in various areas of day-to-day life (such as safety at airports, stations, and in the private sector) as well as in many more specialized domains (e.g. traffic monitoring or safety in petrol, etc). Existing surveillance systems are not yet capable of autonomous analysis of complex events in camera networks. Here, millions of video feeds are not analyzed in real time and cannot be used for crime or terrorism prevention, etc. The tremendous amounts of information of existing surveillance systems are, at best, recorded or handled by human operators. Using state of the art techniques [1], there are still many issues which have to be solved regarding real time, robustness and environment changes. We aim to develop novel methodologies for intelligent and robust object tracking in smart camera networks. Finally, the goal is a framework which is capable to detect complex event occurring in a smart camera network.